Exercise 2: Two Ways This Study Design Deviates From Nielsen's Own Recommendations — Possible Solution
====================================================================

Deviation one: running a single 15-user study instead of three
separate 5-user studies. Nielsen's own real recommendation is three
rounds of 5 users each, with real fixes made and retested between
rounds - a single big study never gets the chance to catch problems
that a redesign itself introduces, since there's no round after it to
test the fix. Fifteen users tested all at once, against the same
un-revised design, is a genuinely different (and less effective) study
design than what the formula was actually built to support.

Deviation two: not splitting participants across the three genuinely
distinct user types. Nielsen's real guidance is that when a product
has several highly distinct user groups, the 5-per-round figure needs
to be split per group (for example, 3-4 users per group when there are
two groups) rather than treated as one shared pool - a handful of
users drawn from a mix of three very different audiences won't
adequately represent any one of them, since problems specific to one
group can easily go undetected if too few of that group's own users
were actually included.

ANSWER: The two real deviations are (1) running one 15-user study
instead of Nielsen's own recommended three separate 5-user rounds with
fixes and retesting between them, and (2) not splitting participants
across the three genuinely distinct user groups, which Nielsen's own
guidance says needs a separate, smaller batch of users per group
rather than one shared pool.

WHY THIS WORKS AS AN ANSWER
------------------------------
This correctly identifies both of the chapter's own explicitly stated
caveats to the 5-user formula and applies each one to the specific
study design described, rather than treating the formula as a single,
unconditional rule.